Putting trust in the US budget federal trust funds and the politics of commitment (theories of institutional design)

PUTTING TRUSTIN THE US BUDGET: Federal Trust Funds and the Politics of Commitment CAMBRIDGE UNIVERSITY PRESS ERIC M... List of ®gures page viiiList of tables ix 1 Introduction: trust f

PUTTING TRUST

IN THE US BUDGET: Federal Trust Funds and

the Politics of

Commitment

CAMBRIDGE UNIVERSITY PRESS

ERIC M PATASHNIK

In the United States many important programs are paid from trust funds.

At a time when major social insurance funds are facing insolvency, thisbook provides the ®rst comprehensive study of this signi®cant yet little-studied feature of the American welfare state Equally importantly, theauthor investigates an enduring issue in democratic politics: can currentof®ceholders bind their successors? By law, trust funds, which get most oftheir money from earmarked taxes, are restricted for speci®c uses.Patashnik asks why these structures were created, and how they haveaffected political dynamics He argues that of®ceholders have used trustfunds primarily to reduce political uncertainty, and bind distant futures.Based on detailed case studies of trust funds in a number of policy sectors,

he shows how political commitment is a developmental process, wherebyprecommitments shape the content of future political con¯icts This bookwill be of interest to students of public policy, political economy, andAmerican political development

eric patashnik is Assistant Professor in the School of Public Policyand Social Research at the University of California, Los Angeles He waspreviously Assistant Professor of Political Science and Lecturer in Law atYale University He has also been a Research Fellow at the BrookingsInstitution

Series Editor

Robert E Goodin

Research School of Social Sciences

Australian National University

Advisory Editors

Brian Barry, Russell Hardin, Carole Pateman, Barry Weingast,

Stephen Elkin, Claus Offe, Susan Rose-Ackerman

Social scientists have rediscovered institutions They have been ingly concerned with the myriad ways in which social and political institu-tions shape the patterns of individual interactions which produce socialphenomena They are equally concerned with the ways in which thoseinstitutions emerge from such interactions

increas-This series is devoted to the exploration of the more normative aspects ofthese issues What makes one set of institutions better than another? How, if

at all, might we move from the less desirable set of institutions to a moredesirable set? Alongside the questions of what institutions we would design,

if we were designing them afresh, are pragmatic questions of how we can bestget from here to there: from our present institutions to new revitalised ones.Theories of institutional design is insistently multidisciplinary and inter-disciplinary, both in the institutions on which it focuses, and in themethodologies used to study them There are interesting sociological ques-tions to be asked about legal institutions, interesting legal questions to beasked about economic institutions, and interesting social, economic and legalquestions to be asked about political institutions By juxtaposing theseapproaches in print, this series aims to enrich normative discourse sur-rounding important issues of designing and redesigning, shaping andreshaping the social, political and economic institutions of contemporarysociety

Other books in this series

Robert E Goodin (editor), The Theory of Institutional Design

Brent Fisse and John Braithwaite, Corporations, Crime, and AccountabilityItai Sened, The Political Institution of Private Property

Bo Rothstein, Just Institutions Matter

Jon Elster, Claus Offe and Ulrich Preuss, Institutional Design in Communist Societies: Rebuilding the Ship at Sea

Post-Mark Bovens, The Quest for Responsibility

Geoffrey Brennan and Alan Hamlin, Democratic Devices and Desires

FOR AND ON BEHALF OF THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE

The Pitt Building, Trumpington Street, Cambridge CB2 IRP

40 West 20th Street, New York, NY 10011-4211, USA

477 Williamstown Road, Port Melbourne, VIC 3207, Australia

http://www.cambridge.org

© Eric M Patashnik 2000

This edition © Eric M Patashnik 2003

First published in printed format 2000

A catalogue record for the original printed book is available

from the British Library and from the Library of Congress

Original ISBN 0 521 77174 9 hardback

Original ISBN 0 521 77748 8 paperback

ISBN 0 511 01678 6 virtual (netLibrary Edition)

Anne and Bernard Patashnik

List of ®gures page viiiList of tables ix

1 Introduction: trust funds and the politics of commitment 1

2 Political transaction costs, feedback effects, and

9 Barriers to trust fund adoption: the failed cases of energy

security and lead abatement 173

10 Conclusions: the structure and normative challenges

of promise-keeping 188

vii

3.1 General fund vs trust fund taxes, ®scal years 1940±1999 page 433.2 General fund vs trust fund excise taxes, ®scal years 1940±1999 444.1 Pay-as-you-go ®nancing of Social Security, ®scal years

4.2 Annual surplus or de®cit and balance in the Social SecurityTrust Fund as a percent of GDP, ®scal years 1937±1999 875.1 Financial history of HI Trust Fund, ®scal years 1966±1996 1015.2 Premiums as a percent of Part B income, ®scal years

2.1 Four rationales for trust fund ®nancing page 272.2 Diffusion of trust funds across policy sectors 282.3 Mediating effects of trust ®nancing 312.4 Strategies for shaping the future credibility of already

established trust funds 373.1 Signi®cant federal tax legislation: Truman to Reagan

administrations 423.2 Growth rates and stability of selected categories of

federal tax revenue, ®scal years 1948±1997 463.3 Summary of hypotheses 503.4 Determinants of total, general fund, and trust fund taxes,

®scal years 1948±1997 523.5 Determinants of general fund and trust fund excise taxes,

®scal years 1957±1997 533.6 Determinants of top income, current Social Security, and

ultimate Social Security tax rates, 1947±1996 544.1 Social Security surpluses and budget de®cits, 1985±1998

in billions of dollars 925.1 Number of years from Medicare Hospital Insurance

Part A Trustees' projection until insolvency 1036.1 Federal gas tax increases, 1956±1993 1286.2 Average trust fund spending as a percent of annual taxes

and income, ®ve-year periods between ®scal years

7.1 Standard deviation of annual real percentage change in air

ix

transportation budget during the decades immediately beforeand after trust fund creation 1438.1 Budget Authority for Superfund and EPA operations

during early 1980s 16410.1 Trust fund resources by federal missions, ®scal year 1995 192

The topic of this book ± the role of trust funds in American nationalbudgeting ± lies at the intersection of public policy and political science.This is not an accident My graduate training at Berkeley was in both

®elds At Berkeley's Graduate (now Goldman) School of Public Policy, Istudied the ef®cient and equitable design of public policies After Ireturned to Berkeley for doctoral studies in political science (having spenttwo years as a legislative aide in Washington), I became interested in thehistorical and institutional context in which policymaking unfolds, and inthe impact of past choices on present and future options My interest infederal trust funds was originally stimulated by my inability to makesense of contemporary political debates over Social Security I noticedthat whenever lawmakers discussed the relationship between SocialSecurity and the federal budget, they focused on the spending and income

¯ows of the Social Security Trust Fund But what was the signi®cance ofthis arcane ®scal device? If, as I soon discovered, the Social Security TrustFund was not money under the mattress, what governance roles did thedevice perform? Answering that question took me rather far a®eld ± back

to Social Security's adoption and very early development, and to otherpolicy sectors, such as transportation, where the trust fund mechanism isalso employed The common thread was the attempt by current policyactors to put future budget actors under obligation A study of the trustfund structures in the federal budget thus offered the chance to explorethe US government's performance as promise-keeper

Of course, promises must be paid for As I began this project in late

1993, America's social insurance system was coming under scrutiny Ifanything, the political debate has only intensi®ed in the ensuing years As

xi

the baby boomers approach retirement, massive governmental bills arecoming due The prospect of trust fund ``insolvency'' focuses attention onthe need for change, yet of®ceholders are reluctant to break faith withthose who have paid into the current system This book does not evaluateparticular reform options, but its detailed case studies of the origins anddevelopment of the major trust funds provide essential background for

an understanding of today's debates

One of the most pleasant aspects of completing this book is that I can

®nally thank in print the many people and organizations that have helped

me along the way It seems ®tting to begin by expressing my deepgratitude to Professor Eugene Bardach, whose late-night phone callconvinced me to enter Berkeley's graduate program in public policy.Although Gene was not an of®cial member of my dissertation committee,

he has been a superb mentor ± and treasured friend ± ever since Genealso provided extremely helpful comments on several chapters during my

®nal push to ®nish this book

When I migrated across the Berkeley campus from the GoldmanSchool to the Political Science Department, I found an excellent adviser

in Bruce Cain His encouragement, insight, and wit made thesis writingfar more enjoyable than it would otherwise have been I learned atremendous amount from Bruce about how to make the dif®culttransition from student to scholar John W Ellwood shared with me hisextensive knowledge of American national budgeting His stubbornquestions forced me to clarify my arguments; and Henry Brady was a veryhelpful third-reader Finally, I would be remiss if I failed to acknowledgethe considerable in¯uence on my intellectual development of the lateAaron Wildavsky By the time I decided to write a dissertation on abudgeting topic ± one that Aaron had identi®ed in a brief passage asworthy of scholarly attention ± he had passed away But earlier in mygraduate career I did get the chance to take his unforgettable seminar onpolitical culture I feel privileged to have known him

The Brookings Institution afforded me a Research Fellowship in theGovernmental Studies Program, an ideal setting in which to completemost of the research for this book Kent Weaver made signi®cantcontributions to this study He took the time to read both my thesis andsome early conference papers, and offered extremely helpful, detailedsuggestions for strengthening my analysis I also received valuable advicefrom Sarah Binder, Allen Schick, Tom Mann, and Joe White Also I amindebted to my fellow graduate students Gary McKissick, Carolyn Wong,and Julian Zelizer both for their many intellectual contributions to thisproject and for their companionship

Over the years, Julian, a policy historian who shares my interest in the

politics of funding modern American government, has read and rereadcountless versions of my work From my initial efforts to construct anargument to my ®nal revisions, he has improved my thinking withpenetrating insights of his own Without his enormous help, this bookwould look much different.

Since coming to Yale in 1996, I have found a wonderful home in theInstitution for Social and Policy Studies Donald P Green, the Institute'sDirector, created a tremendously stimulating environment that aided mywork in signi®cant ways I would like to thank the Institute's talentedstaff, especially Pam LaMonaca, for all their support I also wish toexpress my appreciation to Anthony Kronman, Dean of the Yale LawSchool, for providing me research support over several summers And I

am happy to give my special thanks to my Yale colleagues and goodbuddies Martin Gilens and Alan Gerber, whose friendship and lunch-timeconversations have kept me grounded and (reasonably) sane

Paul Pierson gave me excellent advice at a critical juncture in thisproject's intellectual development I also wish to thank the scholars whoreviewed this book for Cambridge University Press Their criticisms andsuggestions led to many substantive and stylistic improvements As ithappened, my most meticulous and insightful reader chose not to hideher identity This gives me the opportunity to express my profoundappreciation to Martha Derthick for giving me the bene®t of herexceptional knowledge about American national government Martha'ssage advice helped me bring out the potential of this book None of thesescholars should be blamed for the remaining defects or mistakes; I alone

am responsible

I thank the Academy of Political Science for allowing me to use greatlyrevised material from an essay of mine published in Political ScienceQuarterly, vol 112, no 3 (1997) Many others also deserve thanks: JackCitrin, John Cogan, Cathy Cohen, Tom Cuny, Coco Gordon, MichaelGraetz, Colleen Grogan, Robert Katzmann, Al Klevorick, Jessica Korn,Martin Levin, Ted Marmor, Jerry Mashaw, David Mayhew, MathewMcCubbins, Roy Meyers, Pietro Nivola, Nelson Polsby, Robert Re-ischauer, Susan Rose-Ackerman, Irene Rubin, Mark Schlesinger, StaceySchoenfeld, Ian Shapiro, Stephen Skowronek, Rogers Smith, FredThompson, Dan Tuden, Ben Wildavsky, and Ray Wol®nger I conducted

a number of interviews during the course of my research, and I wouldlike to thank the congressional staff members, lobbyists, and executiveof®cials who took the time to speak with me

Robert Goodin has been an ideal series editor He was enthusiasticabout the project from the start, offered many useful suggestions forstrengthening the manuscript, and was patient with me during the long

process of revisions John Haslam guided the book (and me) through theediting process with a steady hand And I received outstanding copy-editing, indexing, and production assistance from Anne Rix, ShirleyKessel, and Robert Whitelock.

This book is affectionately dedicated to my parents, Anne and BernardPatashnik They have believed in me even when I have doubted myself I

am deeply grateful for their love and support When authors acknowledgetheir children in pages such as these, they often note the welcomedistractions the kids offered from the hard chore of writing My sonsMichael and Josh provided many pleasant diversions, of course, but theyalso let me know, in their own way, that they are proud to have a dadwho teaches and writes books For this, and for the many other preciousgifts they have given me, my heartfelt thanks

My largest personal debt is to my wife Debbie Gordon Since our days

as graduate school classmates in Berkeley (thanks for that phone call,Gene!), Debbie has supported me in every way possible I trust she knowshow much our commitment to one another ± and our life together ±means to me

is central to both party and regime building Yet, while promises arerooted in the imperatives of democratic life, they nonetheless pose aserious political dilemma for democracy Without the ability to endowpolicies with durability, of®ceholders cannot shape the future of thepolity If every governmental promise is written in stone, leaders willeventually lose the capacity to control the present How politiciansmanage and manipulate this fundamental tension between commitmentand ¯exibility is the subject of this book.

As the empirical material for this investigation, the book explores theorigins and evolution of an important yet little-studied institutionalarrangement ± trust funds in the United States national budget Majorexamples include the Social Security Trust Fund, the Medicare HospitalInsurance Trust Fund, and the Highway Trust Fund In contrast togeneral revenues, which are available for the general purposes of govern-ment, trust funds are ``restricted by law to designated programs or uses.''1

The funds obtain most of their revenues from speci®c earmarked taxes(e.g., payroll taxes and gasoline taxes).2Certain trust funds, however, also

1 Allen Schick, The Federal Budget: Politics, Policy, Process (Washington, DC: The Brookings Institution, 1995), 14.

2 This book focuses on trust funds and earmarked taxes in the US national budget Many state budgets also contain trust funds, but they are not speci®cally examined

in this book.

1

receive transfers from within the budget, such as interest payments intothe Social Security Trust Fund By 1995, the more than 150 trust funds inthe US budget comprised almost 40 percent of total federal revenues(excluding internal transfers), up from less than 10 percent in 1950.3

One might think that federal trust funds are an arcane subject, ofinterest only to government accounting freaks Nothing could be moremistaken Trust funds generally embody long-term political commit-ments The trust fund device is meant to provide assurance that policypromises, once made, will be kept ``Federal trust funds,'' writes theCongressional Research Service, ``typically have been established forprograms that have very long-term purposes.''4Many trust funds werefounded with an explicit understanding that ``in exchange for the public'spaying certain taxes or premiums, the government would commit itself to

®nance some activity.''5At times, trust funds have been seen as a vehiclefor building large reserves in order to ``prefund'' future governmentspending In general, however, federal trust funds have been maintained

on a ``pay-as-you-go'' basis, with current taxes used to support currentbene®ts The existence of trust fund ®nancing, however, is still meant tomake long-term promises stick

If trust funds work as intended (something that obviously cannot beassumed), they narrow the ¯exibility of future of®ceholders to allocatebudget resources ± that is, to exercise public authority ± as they see ®t.Whether such efforts by current of®ceholders to tie the hands of theirsuccessors can ever be normatively acceptable, or even successful, in ademocratic polity has long been the subject of debate Many democraticthinkers argue that it is immoral to bind the future ``[E]very age andgeneration must be as free to act for itself, in all cases, as the ages and

3 Some 179 budget accounts of®cially designated as trust funds existed on the books

of the US Treasury in 1995 This ®gure exaggerates the true number of signi®cant trust fund programs, however First, some trust fund programs have multiple accounts Second, the ®gure includes more than 35 minuscule trust funds

established to carry out a conditional gift or bequest (Believe it or not, some people actually voluntarily donate money to the US government.) Virtually every executive department has at least one conditional gift trust fund These funds have little political signi®cance and are not examined in this study See General Accounting Of®ce, ``Budget Account Structure: A Descriptive Overview,'' GAO/ AIMD-95-179 (Washington, DC: Government Printing Of®ce, 1995).

4 David Koitz, Dawn Nuschler, and Philip Winters, ``Federal Trust Funds: How Many, How Big, and What Are They For?'' CRS Report for Congress, Updated August 30, 1996 (Washington, DC: Congressional Research Service, 96±686 EPW), 2.

5 Ibid.

generations which preceded it,'' wrote Thomas Paine.6 In Paine's view,democracy is meaningless if current leaders are blocked from choosingtheir own path Thomas Jefferson largely agreed with this position WhileJefferson endorsed limited constitutional protection of basic rights, hebelieved that of®ceholders had no right to legislate for the distant future.Accordingly, he insisted that all laws and institutional arrangements mustlapse at set intervals National plebiscites would then be held to determinethe new form of government.7The great nineteenth-century British legalscholar Alfred Dicey argued that attempts to bind the future were not somuch immoral as futile ``That Parliaments have more than once intendedand endeavored to pass Acts which should tie the hands of theirsuccessors is certain, but the endeavor has always ended in failure.''8

According to Dicey, a ``sovereign power cannot, while retaining itssovereign character, restrict its own powers by any particular enact-ment.''9

The creation of trust funds, and the politics surrounding theiroperation, thus provides an excellent setting in which to explore thelimits and possibilities of statutory commitment in democratic politics.Why have politicians created trust funds for some programs but notothers, and how has the preexistence of trust funds shaped ensuing policyoutcomes? What actors have participated in trust fund decisions and whyhave some trust fund arrangements proved more stable than others? Canpoliticians ``undo'' inherited commitments when they become incompat-ible with present needs? Finally, what are the normative challengesassociated with dedicated funding and how can the future use of the trustfund instrument in public budgeting be improved?

The need for this study

A variety of governmental institutions ranging from civil service tobureaucratic structure can be seen as political commitment devices.10

Most political scientists have directed their attention to written

constitu-6 Quoted in Stephen Holmes, ``Precommitment and the Paradox of Democracy,''

in Jon Elster and Rune Slagstad, eds., Constitutionalism and Democracy

(Cambridge: Cambridge University Press, 1988), 200 Holmes provides a

provocative normative discussion of the issue.

tions, which determine the basic sources of authority in a given polity Ingeneral, constitutional provisions are indeed quite resistant to change.11

Yet the same basic tension between commitment and ¯exibility also arises

in the statutory realm To be sure, the commitments embodied inordinary statutes are generally much easier to reverse than those written

in constitutions But statutory commitments nonetheless often provequite durable and long lasting In an era of big government, the impact ofpast legislative promises on contemporary dynamics simply cannot beignored

Indeed, as the modern welfare-administrative state matures, and therange and sheer number of commitments on the statute books expands,with new programs periodically added to a generally stable base of priorobligations, the very nature of governance changes Increasingly, of®ce-holders in all the major industrialized democracies ®nd that they areconstrained by the legacy of previous administrations As Richard Roseand Philip Davies pointedly argue,

The familiar maxim to govern is to choose is reductionist in the extreme Itimplies that government is carried out by individual decisionmakers whohave as much freedom of choice as an individual in a shopping malltrying to decide whether to have a pizza or an ice cream cone Thestatutory commitments of a newly installed of®cial are not a menuspecifying what an individual might choose but a description of what apolicymaker is committed to do Like it or not, each new arrival in of®cemust recognize that to govern is to inherit.12

The legacy of prior commitments is evident in every policy domainbut perhaps nowhere more so than in US budget politics, wheredemographic and economic factors combine with unusually intricatepatterns of institutional design to severely restrict the formal discretionavailable to of®ceholders Annually appropriated spending ®nancedthrough general tax revenues ± the type of spending most susceptible tothe control of incumbent politicians ± today comprises one-third of the

US budget, down from two-thirds a generation ago.13 According to one

11 Of course, the practical meaning of constitutional provisions is by no means immune to change For a provocative argument that ``higher lawmaking'' can occur even without formal constitutional amendment, see Bruce A Ackerman,

We the People (Cambridge, MA: Harvard University Press, 1991).

12 Richard Rose and Philip L Davies, Inheritance in Public Policy: Change without Choice in Britain (New Haven: Yale University Press, 1994), 1±2.

13 Robert D Reischauer, ``The Unful®llable Promise: Cutting Nondefense

Discretionary Spending,'' in Robert D Reischauer, ed., Setting National Priorities: Budget Choices for the Next Century (Washington, DC: Brookings Institution, 1997).

policy expert, if this trend is not halted, Americans will soon face apolitical future that is little more than the ``preprogrammed outcome ofpromises made by past elected of®cials.''14

Of course, all budget items involve commitments for the future Butsome commitments in the budget are reinforced by mechanisms explicitlydesigned to narrow future of®ceholders' freedom of choice Three suchmajor devices exist in the American national budget, of which the trustfund mechanism is one The two others are entitlements (provisions oflaw that mandate certain payments to eligible persons) and indexation(automatic program adjustments for in¯ation).15The three devices some-times occur in conjunction with one another For example, many of thebig indexed entitlement programs (e.g., Social Security) are paid fromtrust funds But trust funds are also used to ®nance so-called discre-tionary programs, such as highway building Each of the devices deserves

to be the focus of study in its own right, because each ``may tie legislators'hands in different ways.''16Despite trust funds' massive scope and theirimportant role in ®nancing core US domestic programs, however, nogeneric study of the trust fund device exists.17

Trust funds merit attention on both substantive and theoreticalgrounds Programmatically, the durability of government programs oftenturns on continuity of funding Indeed, the entitlement status of anumber of social programs without a dedicated funding source hasrecently been either seriously challenged (food stamps, Medicaid) orrepealed (Aid to Families with Dependent Children) Trust funds them-selves have moved to the center of recent debates over U.S social policybecause the Social Security and Medicare trust funds are reportedlyheading toward ``bankruptcy.'' How to keep these programs solvent forretiring baby boomers is a hot-button issue Any contemporary under-

14 Eugene Steuerle, ``Discretion to Do the Right Things,'' The Washington Post, May

18, 1998, A17.

15 Joseph J Cordes, ``How Yesterday's Decisions Affect Today's Budget and Fiscal Options,'' in C Eugene Steuerle and Masahiro Kawai, eds., The New World Fiscal Order: Implications for Industrialized Nations (Washington, DC: The Urban Institute, 1996), 95±116.

16 C Eugene Steuerle and Masahiro Kawai, ``The New World Fiscal Order:

Introduction,'' in C Eugene Steuerle and Masahiro Kawai, eds., The New World, 8.

17 By contrast, policy scholars have carefully analyzed the politics of entitlements and indexation Kent Weaver has done some of the best work on these topics See

R Kent Weaver, ``Controlling Entitlements,'' in John E Chubb and Paul E Peterson, eds., The New Direction in American Politics (Washington, DC: The Brookings Institution, 1985); and R Kent Weaver, Automatic Government: The Politics of Indexation (Washington, DC: The Brookings Institution, 1988).

standing of the American welfare state must pay close attention to trustfunds and earmarked taxes.

US budget trust funds are theoretically intriguing because theirsigni®cance as a commitment device is not obvious The promises under-lying public trust funds are not subject to an external enforcementmechanism Moreover, as presently constituted, federal trust funds do notcumulate real wealth When a trust fund takes in more than it pays out,the cash goes into the general Treasury, and the trust fund is creditedwith a non-marketable federal security These securities are backed by thefull faith and credit of the United States,18 earn interest at competitiverates of return,19 and are subject to the legal debt ceiling set byCongress.20Moreover, the securities add to national savings to the extenttrust fund surpluses reduce the level of borrowing the federal governmentwould otherwise incur But trust fund reserves do not themselvesconstitute economic wealth The reserves are ``claims on the Government,not for the Government.''21When the time comes to make good on trustfund spending promises, the government must do what it ordinarily does

to ®nance programs ± raise taxes, reduce other expenditures, or increasepublic borrowing

Given all this, many observers assert that the US government's trustfunds are ``bogus.''22A former top budget of®cial states, ``Remember how

18 General Accounting Of®ce, ``Financial Audit: 1997 Consolidated Financial Statements of the United States Government,'' GAO/AIMD-98±127

(Washington, DC: Government Printing Of®ce, March 1998), 8.

19 Because of differences in statutory language, the interest rates vary slightly among individual trust funds See Congressional Budget Of®ce, ``Federal Debt and Interest Costs'' (Washington, DC: Government Printing Of®ce, May 1993a),

of 1995, when Treasury Secretary Robert Rubin disinvested two civil service retirement funds See Andrew Taylor, ``Rubin's Footwork Frustrates GOP,'' Congressional Quarterly Weekly Report, December 16, 1995, 3793 On the legal issues, see Thomas J Nicola and Morton Rosenberg, ``Authority to Tap Trust Funds and Establish Payment Priorities if the Debt Limit is Not Increased,'' CRS Report for Congress (Washington, DC: Congressional Research Service,

November 9, 1995, 95±11109 A).

21 David Koitz, Dawn Nuschler, and Philip Winters, ``Trust Funds and the Federal De®cit,'' CRS Report for Congress (Washington, DC: Congressional Research Service, February 26, 1990, 90±106 EPW), 4.

22 ``Trust Them,'' The Wall Street Journal, May 14, 1993, A10.

awful it was when you realized there wasn't a Tooth Fairy or a SantaClaus? Well, brace yourself for another rude awakening After working inthe bowels of federal budgeting for two years, I'm here to tell you thatthere are no trust funds.''23``The federal budget is full of trust fundsthat deserve neither half of the name They contain no funds andbecause they mainly exist on paper, they don't inspire much trust, either;they shouldn't anyway,'' editorializes The Washington Post.24 ``[I]n anysense that matters, the funds do not exist,'' echoes The Economist.25

Two arguments are being made here The ®rst is manifestly wrong,while the second is largely correct but grossly incomplete The erroneousclaim is that the creation of a federal trust fund has no bearing on thepolitics of resource allocation Certainly, clientele groups believe other-wise; and they back their beliefs by devoting resources to lobbying fortrust fund status.26 Why do the forecasts of the Social Security andMedicare Boards of Trustees receive such enormous political attention?27

Why have trust fund taxes grown so much more rapidly than generalfund taxes since the 1950s?28The evidence presented in this book suggeststhat federal trust funds do make a difference, just as many political actorsthink they do As a former Public Trustee of the Social Security andMedicare Trust Funds argues, ``While in recent years there has beenconsiderable criticism of trust funds as being illusory, such allegationsare not well informed and ignore both the legal and the administrativerealities surrounding [them].''29

A more credible, yet still incomplete, argument is that federal trustfunds are not what they seem to be ± exact replicas of trust fundsmaintained in the private sector In the private sector, a trust is a ®duciaryrelationship in which one person (the trustee) holds property for the

23 Matthew Miller, National Public Radio Commentary, October 16, 1995 I thank

Mr Miller for sending me a transcript of his remarks.

24 ``Trust, but Verify,'' The Washington Post, December 1, 1997, A24.

25 ``Put not your trust in Congress,'' The Economist, November 11, 1989, 56.

26 On lobbying for trust funds, see Lawrence J Haas, ``Paying As You Go,'' National Journal, October 22, 1988, 2644±8; see also Roy T Meyers, Strategic Budgeting (Ann Arbor: University of Michigan Press, 1994).

27 On the attention given to projections of the ``bankruptcy'' of the Medicare Trust Fund, see David Rosenbaum, ``Gloomy Forecast Touches Off Feud on Medicare Fund,'' The New York Times, June 6, 1996, A1.

28 On the tremendous increase in trust fund taxes, see John F Cogan, ``The Dispersion of Spending Authority and Federal Budget De®cits,'' in John F Cogan, Timothy J Muris, and Allen Schick, eds., The Budget Puzzle (Palo Alto: Stanford University Press, 1994).

29 Stanford G Ross, ``Institutional and Administrative Issues,'' in Eric R Kingson, and James H Schultz, eds., Social Security in the 21 st Century (New York: Oxford University Press, 1997), 231.

bene®t of another (the bene®ciary) The trustee's obligation under trustlaw is to manage the assets of the trust property ``solely in the interests ofthe bene®ciary.''30 Private trusts originated in the late Middle Ages as adevice for transferring wealth within the family The purposes of trustshave since mushroomed Trusts are now used in a wide variety ofcommercial settings Interestingly, private sector trusts remain a ``un-iquely Anglo-American institution,'' having never taken root in Con-tinental legal systems.31 In both the United States and the UnitedKingdom, political leaders often appeal to the trust analogy whenarticulating visions of good government British civil servants, forexample, are said to hold ``positions of trust under the Crown.''32``Publicof®ce is a public trust'' was used by Grover Cleveland as the motto for hisadministration Such language is meant to convey not that those whooccupy positions of responsibility hold legal title to their of®ces, only thatof®cials are properly held to a standard of good faith.

The analogy to private trusts is much closer in the case of trust funds

in the American national budget Until fairly recently of®cial US ment documents in fact de®ned trust funds as being ``held in a ®duciarycapacity,'' the theory being that the money was not really ``owned'' by thefederal government.33 Unlike their private sector counterparts, however,the overwhelming majority of federal trust funds are not based on a true

govern-®duciary relationship.34 While the creation of a government trust funddoes involve a legal commitment to use the money for speci®ed purposes,Congress has the right to ``unilaterally alter'' the tax rates and bene®tlevels of trust fund programs by changing existing law.35 This applieseven to those trust funds that ®nance entitlement programs like Social

30 George Gleason Bogert and George Taylor Bogert, Handbook of the Law of Trusts,

®fth edition (St Paul, Minn.: West Publishing, 1973), 2.

31 John H Langbein, ``The Contractarian Basis of the Law of Trusts,'' Yale Law Journal, 625, 1995, 632±43, at 669.

32 Although the trust analogy is often invoked in British government, use of tax earmarking is relatively rare See Barry Bracewell-Milnes, ``Earmarking in Britain: Theory and Practice,'' The Case for Earmarked Taxes: Government Spending and Public Choice (London: Institute of Economic Affairs, 1991).

33 See Tax Foundation, Federal Trust Funds: Budgetary and Other Implications (New York: Tax Foundation, 1970), 5.

34 The federal government does have a ®duciary responsibility for several trust funds, including assets held in trust on behalf of American Indian tribes In fact, however, Indian tribal funds have a sad history of gross mismanagement by the government See Rochelle L Stan®eld, ``Why Indian Trust Funds are in

Disarray,'' National Journal, May 2, 1992, 1062.

35 General Accounting Of®ce, ``Budget Issues: Trust Funds and Their Relationship

to the Federal Budget,'' GAO/AFMD-88±55 (Washington, DC: Government Printing Of®ce, 1988a), 6.

Security and Medicare, in which eligible persons have a legal right tobene®t payments.

Viewed in another light, however, private and public trusts do share anessential feature Both can be seen as a kind of contract ± an agreement to

do something in the future To be sure, private trusts have seldom beendescribed as deals Yet as Yale legal scholar John H Langbein points out,most private trusts are in fact ``functionally indistinguishable from themodern third-party-bene®ciary contract.''36 The contract involves anagreement between the person who creates the trust (the settler) and thetrustee about how the trust will be managed for the bene®ciary Mostprivate trusts contemplate long duration.37 Contrary to popular belief,private trusts almost never attempt to anticipate every possible futurecontingency Trustees usually possess the discretion to make decisions ascircumstances change, subject to a legal duty of good faith

Many federal trust funds can also be seen as involving a kind ofcontracting behavior ± a commitment from the government to constitu-ency groups that particular governmental activities will be funded in acertain way.38Political actors themselves often describe government trustfunds in precisely these terms For example, Bud Shuster (R-PA), thechairman of the House Public Works Committee, has stated that theHighway Trust Fund constitutes ``nothing less than a contract betweenthe government and the American traveling public.''39Some of the verysame attributes that make the trust device attractive in private settingsalso commend it to political actors An example is the segregationrequirement, which mandates that resources of the trust be earmarked Inthe private sector, this means the property of the trust must be sharplydistinguished from the trustee's own property In the public sector, itmeans that trust fund income must be accounted for separately fromgeneral tax receipts, thus allowing the program's status to be easilyinspected by interested parties.40 Keeping track of speci®c, long-term

36 Langbein, ``The Contractarian Basis of the Law of Trusts,'' 627 Langbein excludes from this account both charitable trusts and constructive trusts imposed coercively to prevent unjust enrichment.

37 Ibid., 654.

38 For a roughly analogous discussion of the similarities and differences between

``social insurance'' and ``private insurance,'' see Robert J Myers, Social Security (Bryn Mawr, Pennsylvania: McCahan Foundation, 1975), 13.

39 On Shuster's conception of the trust funds as contracts, see Kirk Victor, ``Trust Me,'' National Journal, March 11, 1995, 607±11.

40 Note the segregation mandate does not prohibit the government from

commingling the actual cash generated by earmarked and general fund taxes There is no separate drawer in the Treasury labeled the Social Security Trust Fund.

budget promises can be quite dif®cult when various programs are rolledtogether into a single agency budget Separate fund accounting reducesthe political transaction costs of monitoring compliance with particulardeals.

These common elements notwithstanding, there are major differencesbetween public and private trust funds As I mentioned, public trustfunds generally do not involve a contractual relationship from a strictlylegal standpoint because their provisions can be changed by Congress Inthe case of Social Security, for example, the Supreme Court has ruled thatthe interest of workers in their pensions ``cannot be soundly analogized tothat of the holder of an annuity, whose right to bene®ts are bottomed onhis contractual premium payments.''41 Indeed, the two forms of trustfunds may lead to opposite effects One key advantage of private trusts, forexample, is protection from the risk of insolvency Even if a privatetrustee experiences personal ®nancial losses, bene®ciaries retain theirinterest in the trust Ironically, the situation is just the reverse in federalbudgeting It is precisely dedicated funding that makes the threat ofinsolvency possible As Martin Feldstein observes, ``Social Security is said

to be heading toward bankruptcy only because it uses earmarked taxesand has a trust fund Other federal programs like education and defensehave no earmarked taxes and no trust fund and would therefore never beperceived to be bankrupt.''42

Ultimately, the essential difference between private and public trusts isnot that the former are somehow more ``real'' than the latter The keydistinction is that private trust funds implement private deals, subject toprivate sector enforcement, whereas public trust funds implement socialcontracts, which are subject to political enforcement As Alan Blindercorrectly argues, many economists may call federal trust funds ``®ctions,but they are facts ± because they have standing in law.''43 What requiresscrutiny is why elected of®cials create trust funds, the nature of theunderlying commitments, and how trust fund structures create politicalfacts and shape bene®ciary expectations

Understanding trust fund commitments

Two social science literatures offer limited insights into these issues The

®rst is the empirical literature on policy inheritances This literature

demonstrates that the actions of incumbent administrations are massivelyconstrained by the actions of their predecessors, even though currentpoliticians always retain the formal authority to remake any law theywish.44This research tends to examine the impact of policy inheritances

on the projects of politicians at particular moments, however In sodoing, it directs attention away from both how speci®c policy commit-ments evolve over very long periods of time and how they were crafted inthe ®rst place

Another relevant body of work is the growing game-theoretic literature

on ``credible commitment.'' This research basically treats commitmentmaking as an exercise in self-restraint The central problem is said to beone of ``time inconsistency.'' The problem arises when some movemaximizes an actor's utility before the fact, but is incompatible with hisincentives afterward Under such circumstances, any commitment toperform the act is simply not credible One solution is for the person tomanipulate his future self by making reneging very dif®cult, as in thefamous tale of Ulysses lashing himself to the mast to prevent his beingfatally attracted to the Sirens' song

Credible commitment often works through reputation effects Forexample, if ``a union leader stakes his reputation on his refusal to approve

a contract that includes a reduction in wages, he might thereby gain theupper hand in wage negotiations.''45This reputation for trustworthiness,

of course, will greatly redound to the union leader's bene®t in subsequentrounds of the game Does this logic apply to federal trust funds? Politicalactors certainly do pledge ®delity to trust funds One common vehicle isthrough political party platforms The 1996 Republican platform, forexample, included the following language:

We have a legal and moral responsibility to America's seniors and willcontinue to do everything in our power to ensure that government honorsour commitment to Social Security bene®ciaries, now and in the future

44 See especially the excellent analyses contained in Rose and Davies, Inheritance in Public Policy; and Paul Pierson, Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment (Cambridge: Cambridge University Press, 1994).

45 Donald C Hubin, ``Of Bindings and By-Products: Elster on Rationality,'' Philosophy and Public Affairs, 15, Winter 1986, 82±95 See as well Jon Elster, Ulysses and the Sirens: Studies in Rationality and Irrationality (Cambridge: Cambridge University Press, 1979); Thomas C Schelling, The Strategy of Con¯ict (Cambridge, MA: Harvard University Press, 1960); and Douglass C North and Barry R Weingast, ``Constitutions and Commitment: The Evolution of

Institutions Governing Public Choice in Seventeenth-Century England,'' in Lee J Alston, Thrainn Eggertsson, and Douglass C North, eds., Empirical Studies in Institutional Change (New York: Cambridge University Press, 1996).

We will keep it ®nancially sound and keep politics out of its tion We will work to ensure the integrity and solvency of the SocialSecurity trust funds.46

administra-Although often quite vague and open-ended, the promises contained inparty platforms are far from meaningless.47 Still, it would be a gravemistake to look to party manifestos as a major source of trust fundcredibility In the ®rst place, few trust funds outside of Social Securityever receive a mention in platform planks Moreover, US political partiesare simply too weak to serve as effective guarantors of policy credibility.48

An even more fundamental problem with this perspective is that veryoften the actors who made the original trust fund vows are dead (oranyway well out of politics) by the time the promises come due It is otherpoliticians, who come later, whose hands the trust funds are meant to tie.With trust funds, it is thus often more a case of binding distant futuresthan credible commitments as such.49

At the deepest level, the problem with the credible commitmentsliterature is its assumption of a perfect analogy between individual andcollective commitment.50 Many rational-choice scholars argue that amajor effect of political institutions ± the congressional committee system

is a favorite example ± is to solve the time-consistency problem bypreventing opportunistic actors from reneging on their bargains This, inturn, expands the set of legislative deals that can be struck, fosters politicalexchange, and ultimately widens the range of options the government cantake in promoting social welfare.51

In fact, trust funds do perform this role, at times at least By providing

a linkage between speci®c revenues and outlays, trust funds and marked taxes may allow deals to be struck that otherwise could not This

ear-is in fact part of the story But it ear-is de®nitely not the whole story The

49 I owe this insight to Robert Goodin.

50 On the disanalogies between individual and collective commitment, see John Elster, ``Introduction,'' in Elster and Slagstad, Constitutionalism and Democracy, 8±14; and Stephen Holmes, ``Precommitment and the Paradox of Democracy,''

in ibid., 236±8.

51 For this side of the theoretical story, see Kenneth A Shepsle, ``Discretion, Institutions, and the Problem of Government Commitment,'' in Pierre Bourdieu and James S Coleman, eds., Social Theory for a Changing Society (Boulder, CO: Westview Press, 1992), 245±63.

other side is suggested in an important series of essays on politicalcommitment by Terry Moe.52 According to Moe, political actors craftcommitment devices not only to promote voluntary exchange, or to solvecollective action problems, but also to serve their long-run interests at theexpense of their current and future opponents Fearing their favoredpolicies will be vulnerable to the interventions of subsequent politicians,actors try to entrench their preferences through structural design Theupshot is that even when precommitments are the product of deliberateplanning, and serve the goals of their designers, they do not necessarilypromote social welfare.53

Any normative appraisal of prior commitment, then, should rest on arealistic understanding of democratic politics Unfortunately, this is nottrue of most writing by economists on earmarked taxes and trust funds.The leading opponents of dedicated funding are orthodox public ®nancescholars, who contend that earmarking introduces ``rigidities into bud-geting, preventing resources from being allocated into the highest priorityuses.''54 To public choice school founder James Buchanan, by contrast,tying the government's hands through earmarking and user charges ispotentially a good thing General fund budgeting amounts to a ``tie-in''sale While citizens may have diverse preferences, they face a singledistribution of public services to accept or reject Earmarking allowscitizens to express their opinions on each public good separately, therebypromoting individual welfare.55 Second-generation public choice scho-lars, however, have questioned whether earmarking arrangements canactually live up to their theoretical promise, arguing that of®ceholderswill simply hoard dedicated taxes for their own purposes whenever theyare short on cash.56In short, earmarked funds are too rigid in one model;

in the other, not rigid enough

For all their differences these two political models share certain basic

52 Terry M Moe, ``Political Institutions: The Neglected Side of the Story,'' Journal

of Law, Economics, and Organization, 6 Special Issue, 1990a, 213±53; and Terry

M Moe, ``The Politics of Structural Choice: Toward a Theory of Public

Bureaucracy,'' in Oliver Williamson, ed., Organization Theory From Chester Barnard to the Present and Beyond (New York: Oxford University Press, 1990b), 116±53.

53 Adam Przewroski and Fernano Limongi, ``Political Regimes and Economic Growth,'' Journal of Economic Perspectives, 7, Summer 1993, 51±70, at 67.

54 Thomas F Pogue and L G Sgontz, Government and Economic Choice: An Introduction to Public Finance (Boston: Houghton Mif¯in Company, 1978).

55 James M Buchanan, ``The Economics of Earmarked Taxes,'' Journal of Political Economy, 71, October 1963, 457±69.

56 See Fred S McChesney, Money for Nothing: Politicians, Rent Extraction, and Political Extortion (Cambridge, MA: Harvard University Press, 1997).

defects Both make the mistake of assuming that ``the impact of a givenset of institutions on economic policy can be `read' directly from theinstitutional structure itself.''57In fact, the effects of trust funds in the USbudget are subtle and contingent These effects can only be assessedthrough a careful examination of the historical and institutional context

in which particular funds are situated Another ¯aw is that both modelsrest on questionable assumptions about the dynamics of democraticpolitics The traditional normative public ®nance literature presumes thatgovernment is run by a benevolent dictator ± not necessarily the beststarting point for a pragmatic assessment of institutional design andgovernance in pluralist democracies By contrast, many younger publicchoice scholars posit that political actors care only about extracting

``economic rent'' for themselves Undeniably, material incentives ence politics As I shall demonstrate, however, other factors, includingpolicy inheritances, information, and administrative routines, shape out-comes as well

in¯u-Plan of the book

The heart of the book consists of detailed case studies of ®ve trust fundprograms: Social Security, Medicare, Highways, Airport and Airways, andSuperfund In addition, the book also examines two cases ± EnergySecurity and Lead Paint Abatement ± where new trust funds wereproposed but not adopted The nature of the questions addressed in thisstudy makes a comparative case study approach an appropriate one Thepolitical development of individual trust fund programs must be carefullyexamined in order to identify the objectives each trust fund was designed

to serve, the support it has received from key actors, and the evolution ofthe trust fund over time The number of cases included in the studyrepresents an attempt to strike an acceptable compromise between asample size small enough to permit a reasonably detailed analysis of eachtrust fund, and large enough to provide a conceptual foundation forgeneralizations about the role of trust funds in the political process Insome instances, the trade-off between richness of detail and analytictractability is particularly severe Indeed, entire books could be writtenabout each of the cases in the sample The contribution of this study is

57 For a cautionary statement against making just this mistake, see Colleen A Dunleavy, ``Early Railroad Policy,'' in Sven Steinmo, Kathleen Thelen, and Frank Longstreths, eds., Structuring Politics: Historical Institutionalism in Comparative Analysis (Cambridge: Cambridge University Press, 1992), 139.

not to write a de®nitive history of each policy area but rather to providethe ®rst generic study of the trust fund device.

This book is organized as follows Chapter 2 presents the conceptualframework that will guide the case study accounts The framework draws

on two distinct yet complementary strands of the ``new institutionalism'':the transaction cost approach and historical institutionalism.58 Thetransaction cost approach highlights the fundamental problem of politicalcommitment It emphasizes the role of purposive institutional design inshaping political transaction costs ± the costs of negotiating, monitoring,and enforcing agreements The historical-institutional perspective con-tributes three key insights into a study of trust fund politics First, itemphasizes that institutional arrangements mediate policy outcomes butare never the sole cause of them Second, it holds that policymaking isessentially a developmental process: new structures are often built onpreexisting ones, and governmental commitments may create importantfeedback effects.59Finally, the perspective highlights the role of ideas, notjust interests, in policymaking

Chapter 3 examines the place of trust funds in the larger context ofchanging federal tax regimes and uses regression analysis to compare theresponsiveness of trust funds and general funds to various political andeconomic factors Chapters 4 through 9 present the detailed case studies.The ®nal chapter summarizes the main ®ndings, analyzes variationsacross the case studies, and discusses normative implications

The argument in brief

The trust funds in the US budget are, virtually without exception, theproduct of deliberate intervention They are consciously crafted politicalmechanisms intended by their designers to bind the government to itspromises to the public, its constituents

The examination of trust fund experiences offered in this bookdemonstrates that the trust fund device has been a consequential US

58 While these two strands of the new institutionalism have different points of departure, they have been moving closer together For example, economist Avinash K Dixit's recent work on political transaction costs emphasizes that commitment is an evolutionary process See Dixit, The Making of Economic Policy: A Transaction-Cost Politics Perspective (Cambridge, MA: MIT Press, 1996) Some leading scholars (e.g., Douglass North) are claimed by both camps as their own.

59 For an excellent literature review, see Peter A Hall and Rosemary C R Taylor,

``Political Science and the Three New Institutionalisms,'' Political Studies, 44,

1996, 936±57.

policy instrument, shaping both aggregate patterns of US taxation andthe micro dynamics of particular expenditure programs At the macrolevel, trust fund taxes have generally been more stable than general fundtaxes over the post-war era Explicit cuts in social insurance payroll taxes,and other trust fund levies, have rarely been on the agenda and havenever been seen as vehicles for counter-cyclical ®scal policymaking Bycontrast, there have been numerous legislated reductions in income andcorporate taxes not tied to speci®c programs A focus on trust funds andearmarked taxes thus offers insight into how the federal government hasmaintained its revenue base at 18±20 percent of GDP during an era ofdivided party control and increased mistrust in government.

The trust fund device has been an important mediating factor at thelevel of individual programs Trust fund ®nancing clearly does noteliminate the impact of other political forces, such as struggles amongcontending interest groups But trust funds do in¯uence how thesestruggles play out by distributing procedural advantages, reinforcingsymbols of moral deservedness and blame, and affecting perceptions ofpolitical ®delity and defection A striking ®nding is just how muchattention is paid by politicians and interest groups to trust fundaccounting, and how the legislative dynamics of programs like SocialSecurity and Medicare are affected by seemingly ``arti®cial'' changes intrust fund forecasts This ®nding highlights the importance of evenarcane institutions in structuring the politics of a polity Three other

®ndings may surprise many scholars First, a number of major trust fundsowe their origins as much to ®scal conservatives seeking to safeguard theTreasury as to liberal proponents of big government; second, despite theobvious appeal of trust fund ®nancing to pragmatic politicians, trustfunds have not taken over the entire federal budget, indeed they have attimes been rejected by Congress to preserve budgetary ¯exibility; and,third, while trust fund architects generally seek to depoliticize programs,many trust funds have been politically contentious and ®nanciallyunstable

Political scientists are increasingly recognizing that institutions are aforce not only for political inertia but also for political change Trustfunds are an excellent case in point While trust funds channel contestsover scarce budget resources, they are compatible with program trajec-tories ranging from rapid growth to retrenchment When surpluses andde®cits in the trust funds have emerged, Congress has often enacted

``corrective'' legislation designed to restore the funds to balance Thefuture credibility of preexisting trust fund commitments is not ®xed,however There are ongoing struggles among politicians over the rulesand procedures under which particular trust funds will operate And trust

fund promises can be renegotiated In sum, my analysis suggests thatstatutory precommitment in a democratic polity must be seen not as a

®nal policy outcome, but rather as an evolutionary process in which priorchoices shape but do not eliminate the prospects for change

It should be recognized at the outset that all trust funds are not thesame While trust funds constitute a class, their variations must beexplored Such variations, I contend, are largely the product of two keydimensions of the underlying political commitments: reciprocity andreliance The former has to do with the degree to which the governmentand payers are in a reciprocal exchange relationship and whether payersand bene®ciaries signi®cantly overlap with one another The latterconcerns the degree to which promisees make long-term commitments inthe expectation of the funds' continuation and hence become vulnerable

to the consequences of unreliability by the government The concludingchapter shows that different combinations of reciprocity and relianceshape the political dynamics, and the normative contexts, of govern-mental promising and promise keeping

in the US budget Next, it outlines four reasons for the creation of trustfunds It then describes the main effects of the trust fund mechanism onthe policymaking process, and how current of®ceholders can try toincrease or decrease the credibility of existing trust fund commitments.Finally, the chapter discusses the case selection.

Political transaction costs, trust funds, and government

budgeting

It is helpful to begin a study of trust funds with an examination of thelarger budgetary system of which the trust fund instrument is a part.Public budgeting must be understood not merely as a technical exercise

in resource allocation, but rather as the setting for some of the mostcrucial tasks of a democratic polity: mobilizing revenues, deliveringbene®ts to constituencies, safeguarding the public ®sc These tasksinherently require politicians to enter into commitments both with oneanother and with voters and groups in the larger society Budgeting canthus be seen as a form of contracting behavior, in which the governmentpledges ± not always credibly ± to do certain things rather than others inthe future As the late Aaron Wildavsky observed in his classic book ThePolitics of the Budgetary Process:

18

Viewed in [this] light, a budget may be regarded as a contract Congressand the president promise to supply funds under speci®ed conditions,and the agencies agree to spend the money in ways that have been agreedupon Whether or not the contract is enforceable, or whether or notthe parties actually agree about what the contract purportedly stipulates,

is a matter for inquiry To the extent that a budget is carried out, however,

it imposes a set of mutual obligations and controls upon the contractingparties A budget thus becomes a web of social as well as of legalrelationships in which commitments are made by all parties, and wheresanctions may be invoked (though not necessarily equally) by all.1

As I explained in chapter 1, although budgeting ``contracts'' are notcontracts in a legalistic sense, viewing the budget as a set of implicit socialcontracts directs attention to a number of important questions that mightotherwise not be raised Are all budgetary transactions (e.g., pensions,routine administrative operations, and so on) fundamentally the same, or

do some kinds of commitments entail more dif®cult governance lenges than others? What institutions and bene®ciary expectations doparticular contract-like commitments create? Why do different forms ofbudget structures exist?

chal-Transaction costs in budgeting

Recent work on political transaction costs offers a framework to addressthese questions The transaction cost approach was originally developed

to study private sector organization: primarily ®rms, markets, andcommon law institutions A key puzzle for economists is why ®rms,which are islands of planning and managerial control, exist at all, giventhe obvious advantages of market competition Oliver Williamson sug-gests that markets do many things well but also have serious limitations.2

Speci®cally, markets are poorly equipped to manage transactions acterized by high uncertainty, speci®c investments, and ``sunkenness.''While markets supply high-powered incentives that promote ef®ciency,these incentives invite opportunistic behavior when contracts requireongoing relationships, yet parties have dif®culty monitoring one another

char-If parties fear being exploited, they may not agree to the deal in the ®rstplace Moving such transactions from the market to other organizationalsettings may allow the deal to go forward

1 Aaron Wildavsky, The Politics of the Budgetary Process (Boston: Little Brown, 1964), 3.

2 Oliver E Williamson, The Economic Institutions of Capitalism (New York: Free Press, 1985).

Scholars have recently begun to apply the transaction cost approach tothe study of public administration.3 This is a potentially productiveresearch move since transaction costs are no less pervasive in politics than

in economics, and the level and incidence of transaction costs can have asigni®cant impact on policy outcomes As Howard Frant cautions,however, we cannot make ``full use of these ideas by trying to squeeze thepublic sector into a framework designed for the private sector.''4The key

to making good use of the transaction cost perspective in the study ofgovernment is recognizing the distinctive characteristics of politics Fivefactors will be stressed here:

1 Shaky political ``property rights.'' In the private sector, propertyrights are generally secure because they are held by named indi-viduals Political property rights are far more tenuous because theseclaims attach to public of®ces, whose occupants are subject toelectoral turnover

2 Multiple contractors While economic contracts are typically betweentwo clearly identi®able contractors, political contracts often havemultiple parties (e.g., voters, lobbyists) on at least one side of therelationship.5

3 Ubiquity of compromise As Terry Moe stresses, the decentralization

of US formal institutions makes accommodation to political nents ``a virtual necessity'' in the institutional design process.6 Asthe case study chapters will show, the need to satisfy multiplefactions may lead to the creation of trust fund structures withembedded tensions and contradictions

oppo-4 Endogeneity of political transaction costs Political transaction costsare to a signi®cant degree ``given'' by the characteristics of speci®cgovernmental policies But political transaction costs can also beactively manipulated by of®ceholders seeking to further theirpolitical goals.7 Both the underlying attributes of budgeting pro-

3 See, for example, Douglass North, ``A Transaction Cost Theory of Politics,'' Journal

of Theoretical Politics, 2 (4), 1990, 355±67 (1990); Charlotte W Twight, ``Political Transaction Cost Manipulation,'' Journal of Theoretical Politics, 6, 1994, 189±216; Horn, The Political Economy; Dixit, The Making of Economic Policy; and Howard Frant, ``High-Powered and Low-Powered Incentives in the Public Sector,'' Journal

of Public Administration Research and Theory, 3 (6), 1996, 365±81.

4 Frant, ``High-Powered and Low-Powered Incentives in the Public Sector,'' 365.

5 Dixit, The Making of Economic Policy, 48±9.

6 Terry Moe, ``The Politics of Bureaucratic Structure,'' in John E Chubb and Paul

E Peterson, eds., Can the Government Govern? (Washington, DC: Brookings Institution, 1990c), 327.

7 See Charlotte Twight, ``Political Transaction Cost Manipulation.''

mises and the content of politicians' agendas therefore demandattention.

5 Weak incentives for global ef®ciency Two factors may reduce theglobal ef®ciency of political institutions First, the costs associatedwith political commitment making, such as narrowed ¯exibility anddiscretion, are seldom perfectly internalized to the actors making thedeal; instead, they may be borne in large part by opposing coalitionsand future generations Second, institutions may take on a life oftheir own, making reform dif®cult even when the institutions nolonger ®t their environment To be sure, private transactions canalso be subject to increasing returns and ``path dependence.''However serious these problems are in the private sector, they arefar more severe in public settings.8

These differences notwithstanding, a focus on the strategic ment, and shaping in¯uence, of political transaction costs does offerinsights into public administration What makes application of thetransaction cost perspective to the study of government trust fundsespecially appropriate is that the transaction cost model grows out of thesame intellectual tradition as the theory of budgetary incrementalism.Both incrementalism and the transaction cost approach rest on thebehavioral premise of bounded rationality Incrementalism holds thatelected of®cials must ®nd ways to cope with the overwhelming com-plexity of budget decisions In Wildavsky's classic account, politicians areshown to take last year's budget as given, focusing their attention onchanges at the margin.9 Incrementalism can be seen as an effectivemethod for reducing the search and information costs associated withreaching agreements under conditions of uncertainty

manage-Missing from traditional incrementalist theory, however, is explicitattention to what happens after the initial funding decisions are made.Incrementalism basically focuses on the ex ante transaction costs ofmaking budget decisions But there are also the ex post costs ofmonitoring and enforcing taxing and spending promises through time.One reason these costs arise is that, as the transaction cost perspectiveemphasizes, politicians may behave opportunistically Political promises

8 As developed by Williamson, transaction cost economics presumes that private sector institutions in advanced capitalist societies are generally ef®cient (else they would not exist) For a powerful argument that public sector institutions should not be presumed ef®cient because their persistence typically owes a great deal to path dependence and lock-in effects, see Paul Pierson, ``Increasing Returns, Path Dependence, and the Study of Politics,'' revised version of 1996 APSA paper, April

15, 1997.

9 Wildavsky, The Politics of the Budgetary Process.

not compatible with the incentives of current of®ceholders may bebroken Just as private sector actors may remove from the markettransactions in which continuity is important, so politicians may seek toshelter certain budget promises from the normal jockeying for appropria-tions support ± the closest approximation to a competitive resource

``market'' in American government.10

It is reasonable to ask why politicians would ever care about thedurability of budget commitments Rank-and-®le legislators are oftensaid to have very short time horizons But the time horizons of otherpolicymakers may in fact be relatively long Such actors are probably fewand far between in American politics But they do exist One thinks ofprogram builders like Franklin Roosevelt and Lyndon Johnson, andinstitutional ``budget guardians,'' such as long-time Ways and MeansCommittee chairman Wilbur Mills Alternatively, policymakers maythemselves have relatively short time horizons, yet still push for long-term commitment devices because they wish to curry favor with otherpolitical actors who do tend to care greatly about policy durability, such

as well-organized interest groups.11 As we will see in the case accounts,both of these dynamics help explain the creation of trust funds

Rationales for creating trust funds

From the standpoint of orthodox public ®nance theory, the existence ofgovernment trust funds presents a puzzle Dedicating revenue for speci®cuses prevents policymakers from maximizing social welfare by directingrevenues where they are needed most As one budget expert puts it,

``When you earmark, you are giving up the power of the legislature tomanage the ®scal affairs of the state.''12

10 Given the ubiquity of budgetary incrementalism, it might be assumed that funding outcomes are inherently stable ± so there is no need for precommitment devices like trust funds or entitlements Certainly the vast empirical literature on incrementalism in the US budget demonstrates that cabinet departments and other large governmental organizations experience small funding changes most of the time But budget outcomes at the micro programmatic level are more vulnerable to signi®cant ¯uctuations On the competitiveness of the budgetary process and efforts by spending advocates to craft protective budget structures, see the excellent analysis contained in Meyers, Strategic Budgeting See also Charles Stewart III, Budget Reform Politics: The Design of the Appropriations Process in the House of Representatives (New York: Cambridge University Press, 1989).

11 On political time horizons, see Pierson, ``Increasing Returns, Path Dependence, and the Study of Politics.''

12 William T Pound, executive director of the National Conference of State

This puzzle largely disappears once commitment problems are knowledged The question then becomes what makes trust fund ®nancingmore or less attractive to different actors in various circumstances? Thereare four main reasons for creating trust funds: (1) to make users pay; (2)

ac-to maximize agency budgets; (3) ac-to reduce uncertainty; and (4) ac-tosafeguard the Treasury

Making users pay (``bene®t taxation'')

The most common rationale for trust fund ®nancing is to charge users.Economists call this ``bene®t taxation.'' In the pure version of the model(which contrasts with taxation based on ability to pay), governmentservices would be ®nanced entirely through user fees that re¯ect marginalcosts Charging directly for public services may not always be feasible,however For example, pricing highways according to the wear and tearimposed by any single driver would be an administrative nightmare.Earmarking revenue to a trust fund may offer a ``second-best'' solution,provided the revenue source is properly designed By imposing a speci®cearmarked tax on a product (e.g., gasoline) whose usage is correlated withuse of the publicly provided service, the government can distribute costs

in proportion to bene®ts received

Bene®t taxation offers policymakers several advantages First, it enablesthem to recover the costs of government services in which it is feasible toexclude non-payers It also promotes allocational ef®ciency Taxpayerswill only demand an increase in services if they perceive the bene®ts ofthe service increment to exceed the costs Finally, a user±pay approachpromotes equity because those who do not want the service are notforced to pay for it.13

notwithstanding, neither of these conditions obtains in the Superfund case (discussed in chapter 8).

agencies' budgets in order to obtain more power, higher salaries, andother good things.14Niskanen's original model presumes that the bureau

is a monopoly with private information, and that it has a passivelegislative sponsor But bureaus may face intense competition for scarceresources, and appropriations committees may well be active.15 Underthese more realistic conditions, budget-maximizing agencies may ®ndaccess to a dedicated funding source appealing when it offers protectionfrom rival bureaus and insulation from ordinary mechanisms of ®scalcontrol

Reduction of political uncertainty

A third reason for trust fund ®nancing ± reduction of political tainty ± ¯ows from the premise that political actors are risk averse.16

uncer-Actors here are less concerned with obtaining the largest possibleincreases than with funding stability.17 Uncertainty increases the trans-action costs of promise making Policymakers may use trust funds tomanipulate the ex ante transaction costs of negotiating an acceptablebargain and/or the ex post costs of safeguarding promised bene®t ¯ows toconstituents through time

Trust funds can help seal bargains by allowing collective decisions onboth sides of the budget equation ± expenditures and revenues ± to bereached simultaneously Sometimes the enactment of a program requiresits proponents to support a revenue expansion favored by anothercoalition Under general-fund ®nancing, however, there is often a timelag between taxing and spending decisions, creating the potential foropportunistic behavior Lawmakers may vote for the tax only to see theircolleagues renege on their end of the deal What is required to escape thistrap is a mechanism for assuring parties that commitments will be

14 William Niskanen, Bureaucracy and Representative Government (Chicago: Aldine Atherton, 1971).

15 Gary J Miller and Terry M Moe, ``Bureaucrats, Legislators, and the Size of Government,'' American Political Science Review, 77, 1983, 297±323.

16 On risk aversion, see Daniel Kahneman and Amos Tversky, ``Choices, Values and Frames,'' American Psychologist, 39, 1984, 341±50.

17 Roy Meyers suggests that dedicated funding is usually part of a minimax strategy, designed to avoid the ``most distasteful'' budget outcome ± a large reduction in funding See Meyers, Strategic Budgeting, 138 Some scholars argue that

Niskanen's theory of budget maximization should be re®ned to take into account risk aversion See, for example, Andre Blais and Stephane Dion, ``Are Bureaucrats Budget Maximizers?'' The Budget-Maximizing Bureaucrat: Appraisals and Evidence (Pittsburgh: University of Pittsburgh Press, 1991).

honored Dedicated funding, the linkage between speci®c revenues andexpenditures, may enable compromises to be reached.18

Policymakers may also use trust funds as a mechanism for assuringimplementation of long-term promises The incentive to create trustfunds for this purpose is strongest when the net political bene®ts offunding continuity are higher (to the relevant political actors if notnecessarily to society as a whole) than are the net bene®ts of preservingbudget ¯exibility This condition often holds in public budgeting, but notalways Commitment devices like trust funds, entitlements, and indexa-tion tend to weaken the capacity of current of®ceholders to use theannual budget process as a mechanism to control their bureaucraticagents Where agency problems loom large because the underlyingtransaction requires bureaucrats to be granted considerable discretionover resource allocation, as is the case for general administrativespending, these mechanisms may become less appealing, even whenpoliticians are narrowly self-interested.19

Safeguarding the treasury

A ®nal plausible reason for trust fund ®nancing is to protect the generalTreasury The goals are to prevent de®cit spending and perhaps limitgovernment growth In the United States, the norm of a balanced federalbudget has long carried a symbolic importance far beyond its objectiveeconomic meaning, standing for the absence of corruption, socialharmony, and the preservation of republic government.20 Trust fund

®nancing can help protect the Treasury directly by helping to mobilizenew revenues Earmarked taxes are often easier to pass than general fundlevies.21 Because the objective is ®scal restraint and not allocative

18 See Buchanan, ``The Economics of Earmarked Taxes,'' 377±90; and Charles J Goetz, ``Earmarked Taxes and Majority Rule Budgetary Processes,'' The American Economic Review, 58, 1968, 128±36.

19 Horn, The Political Economy of Public Administration, 85.

20 On the power of the balanced budget norm in American history, see James Savage, Balanced Budgets and American Politics (Ithaca: Cornell University Press, 1988).

21 On the greater acceptance of earmarked taxes in the US budget, see Alice M Rivlin, ``The Continuing Search for a Popular Tax,'' American Economic Review,

79 (2), May 1989, 113±17 The appeal of tax earmarking is also discussed in B Guy Peters, The Politics of Taxation: A Comparative Perspective (Oxford:

Blackwell, 1991) For a stimulating theoretical discussion of the issue of voluntary tax compliance, see Margaret Levi, Of Rule and Revenue (Berkeley: University of California Press, 1988) See also Barry R Weingast, ``The Role of Credible Commitments in State Finance,'' Public Choice, 66, 1990, 89±97.